Development and validation of nomograms for predicting overall survival and cancer-specific survival in elderly patients with locally advanced gastric cancer: a population-based study

Abstract

Objective: To evaluate the multiple factors influencing the survival of elderly patients with locally advanced gastric cancer (LAGC) and develop and validate the novel nomograms for predicting the survival. Methods: The clinical features of patients treated between 2000 and 2018 were collected and collated from the Surveillance, Epidemiology, and End Results (SEER) database and three medical centres in China, and the patients were randomly divided into a training cohort (3494), internal validation cohort (1497) and external validation cohort (841). Univariate and multivariate analyses of the prognostic values were performed to identify independent prognostic factors associated with overall survival (OS) and cancer-specific survival (CSS), and two nomogram models were developed. Harrell’s concordance index (C-index) and calibration curves were employed to assess discrimination and calibration. Decision curve analysis (DCA) and receiver-operating characteristic (ROC) curves were utilized to investigate the clinical usefulness. Results: In the SEER database, the 5-year OS of the patients was 31.08%, while the 5-year CSS of the patients was 44.09%. Furthermore, in the external validation set, the 5-year OS of the patients was 49.58%, and the 5-year CSS of these patients was 53.51%. After statistical analysis, nine independent prognostic factors of OS and CSS were identified, including age, race, tumour size, differentiation, TNM stage, gastrectomy type, lymph node metastasis (LNM), lymph node ratio (LNR) and chemotherapy. The C‐index (approximately 0.7) and calibration curve (close to the optimal calibration line) indicated satisfactory discrimination and calibration of the nomogram. DCA and ROC curves showed that the developed nomogram was superior to TNM stage. Conclusion: The novel validated nomogram could accurately predict the prognosis of individual elderly patients with LAGC and guide the selection of clinical treatment measures.